SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2007.06833 · 2020

Sudo rm -rf: Efficient Networks for Universal Audio Source Separation

arXiv · PDF · Open in the Atlas

Code that ran

10 of 10 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 2 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2007.06833.json · badge: /api/ran/2007.06833.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
etzinis/sudo_rm_rf canonical (classifier label)
classifier basis: census
10 ran · 0 no recorded run
FunctionStatusWhere it lives
capture_init Ran
ran; our contract draft was wrong, not the code
etzinis/sudo_rm_rf/sudo_rm_rf/dnn/models/demucs.py
code served (permissive licence) · get_code("55ffcb5555a13036")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned function. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import functools
import torch

def capture_init(init):

    @functools.wraps(init)
    def __init__(self, *args, **kwargs):
        self._init_args_kwargs = (args, kwargs)
        init(self, *args, **kwargs)
    return __init__


# Create a test class that uses capture_init
class TestClass:
    @capture_init
    def __init__(self, x, y, z=None):
        self.x = x
        self.y = y
        self.z = z


torch.manual_seed(0)

# Set up arguments for capture_init
# capture_init takes one parameter: init (a function/method)
# We'll use a simple init function that takes self and some arguments

def simple_init(self, a, b, c=10):
    self.a = a
    self.b = b
    self.c = c

ARGS = {
    "init": simple_init
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
center_trim Ran
ran on a synthesized input
etzinis/sudo_rm_rf/sudo_rm_rf/dnn/models/demucs.py
code served (permissive licence) · get_code("f89970356c8e9506")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 6) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function center_trim trims a tensor along the last dimension
# It uses:
# - tensor.size(-1) to get the last dimension size
# - reference.size(-1) if reference has a 'size' method, otherwise treats it as a number
# - tensor[..., start:end] slicing

# Create a tensor with a larger last dimension
tensor = torch.randn(2, 4, 10)  # batch=2, middle=4, last_dim=10

# Create a reference tensor with a smaller last dimension
reference = torch.randn(2, 4, 6)  # batch=2, middle=4, last_dim=6

ARGS = {
    "tensor": tensor,
    "reference": reference,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 4, 6) — shape (2, 4, 6) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
create_input_for_model Ran
ran on a synthesized input
etzinis/sudo_rm_rf/sudo_rm_rf/utils/extract_model_performance.py
code served (permissive licence) · get_code("4cebc9ef51f782a1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

ARGS = {
    'batch_size': 2,
    'input_samples': 4,
    'model_type': 'baseline_demucs'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 3, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
create_targets Ran
ran on a synthesized input
etzinis/sudo_rm_rf/sudo_rm_rf/utils/extract_model_performance.py
code served (permissive licence) · get_code("ff9abcf53349bd98")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (2, 3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

ARGS = {
    'batch_size': 2,
    'input_samples': 4,
    'n_sources': 3,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
overlap_and_add Ran
ran on a synthesized input
etzinis/sudo_rm_rf/sudo_rm_rf/dnn/models/dprnn.py
code served (permissive licence) · get_code("e7f1cfe9c1b6a1aa")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 8) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import math

torch.manual_seed(0)

# Create a signal tensor with shape [..., frames, frame_length]
# Using distinct sizes: batch=2, frames=3, frame_length=4
signal = torch.randn(2, 3, 4)

# frame_step must be <= frame_length
frame_step = 2

ARGS = {
    "signal": signal,
    "frame_step": frame_step,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 8) — shape (2, 8) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
pit_loss Ran
ran on a synthesized input
etzinis/sudo_rm_rf/sudo_rm_rf/dnn/losses/sisdr.py
code served (permissive licence) · get_code("c0e42ad7a931f89c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import itertools

torch.manual_seed(0)

# Create small tensors for y and z
# y shape: (batch=2, sources=3, time_steps=4)
# z shape: (batch=2, sources=3, time_steps=4)
y = torch.randn(2, 3, 4)
z = torch.randn(2, 3, 4)

# Parameters for pit_loss
of = 0  # offset parameter
SI = False  # Scale-Invariant flag

ARGS = {
    "y": y,
    "z": z,
    "of": of,
    "SI": SI,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
remove_pad Ran
ran; our contract draft was wrong, not the code
etzinis/sudo_rm_rf/sudo_rm_rf/dnn/models/dprnn.py
pointer only (licence: MIT); the stored copy get_code serves, from yoonsanghyu/FaSNet-TAC-PyTorch, is pointer only (licence: NONE) · get_code("dcd526946b35bfad")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from yoonsanghyu/FaSNet-TAC-PyTorch, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

sdri_loss Ran
ran on a synthesized input
etzinis/sudo_rm_rf/sudo_rm_rf/dnn/losses/sisdr.py
code served (permissive licence) · get_code("8550c5757f51508f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create small tensors for y and z
# y and z are audio signals that will be processed
# They need to support .unsqueeze(0), .mean(-1), and arithmetic operations
# Let's make them 2D or 3D tensors with small dimensions

batch_size = 2
sequence_length = 4

# y is the reference signal
y = torch.randn(batch_size, sequence_length)

# z is the predicted/estimated signal (same shape as y)
z = torch.randn(batch_size, sequence_length)

# of (offset) is a scalar float that gets subtracted from the SDR
of = 0

ARGS = {
    'y': y,
    'z': z,
    'of': of,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
sisdr_loss Ran
ran on a synthesized input
etzinis/sudo_rm_rf/sudo_rm_rf/dnn/losses/sisdr.py
code served (permissive licence) · get_code("467dcd7fbee6a112")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create small tensors for y and z
# y: reference signal, shape (batch, channels, samples)
# z: predicted signal, shape (batch, channels, samples)
y = torch.randn(2, 3, 4)
z = torch.randn(2, 3, 4)

# of: offset parameter (default 0)
of = 0

ARGS = {
    'y': y,
    'z': z,
    'of': of
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
upsample Ran
ran on a synthesized input
etzinis/sudo_rm_rf/sudo_rm_rf/dnn/models/demucs.py
code served (permissive licence) · get_code("bf40d3a8385e7e43")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small input tensor: (batch=2, channels=4, time=3)
# The function expects 3D tensor and accesses x[..., :-1, :] and x[..., 1:, :]
# So we need at least time=2 to have valid slicing, but time=3 gives better coverage
x = torch.randn(2, 4, 3)

# stride parameter: integer that determines upsampling factor
stride = 2

ARGS = {
    "x": x,
    "stride": stride
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 4, 4) — shape (2, 4, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 2 later) and 7 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
capture_init 2 other papers
1 dated earlier, 1 later
2003.01531 (2020-03)
Voice Separation with an Unknown Number of Multiple Speakers
2211.12232 (2022-11)
AERO: Audio Super Resolution in the Spectral Domain
center_trim 2 other papers
1 dated earlier, 1 later
1809.07454 (2018-09)
Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation
2211.08553 (2022-11)
Hybrid Transformers for Music Source Separation
remove_pad 1 other paper
1 dated earlier
1910.14104 (2019-10)
End-to-end Microphone Permutation and Number Invariant Multi-channel Speech Separation

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

In this paper, we present an efficient neural network for end-to-end general purpose audio source separation. Specifically, the backbone structure of this convolutional network is the SUccessive DOwnsampling and Resampling of Multi-Resolution Features (SuDoRMRF) as well as their aggregation which is performed through simple one-dimensional convolutions. In this way, we are able to obtain high quality audio source separation with limited number of floating point operations, memory requirements, number of parameters and latency. Our experiments on both speech and environmental sound separation datasets show that SuDoRMRF performs comparably and even surpasses various state-of-the-art approaches with significantly higher computational resource requirements.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2007.06833")
get_code_for_paper("2007.06833")
have("2007.06833")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2007.06833.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2007.06833.svg)](https://syntology.ai/paper/2007.06833)

Connect an agent — have() is free.